Louis Dufresne

Investigation of multiscale subgrid models for LES of instabilities and turbulence in wake vortex systems

This paper investigates the capabilities of different subgrid scale (SGS)models, including the recent “multiscale” models, for large-eddy simulation (LES),here in a vortex-in-cell (VIC) method, of complex wake vortex dynamics. More specifically, we here consider the multiscale dynamics developing in a counterrotating four-vortex system, that evolves from a simple state to a turbulent state. The various SGS models are tested and compared on this complex and transitional flow. Comparisons are also made with results obtained using a pseudo-spectral method. Energy diagnostics (global and modal) and spectra are provided and used to support the comparisons. A discussion on the applicability of the various models to LES of complex wake vortex flows is made. The multiscale models are seen to be the most appropriate

Experimental Study on the Influence of Dune Dimensions on Flow Separation

AbstractUsing laboratory experiments and a numerical model, this study examined the influence of dune dimensions on the separation of flow. Different conditions, depending on the dune height, material, and angle and flow depth and velocity, were considered. Experiments were conducted in a horizontal flume with one side wall being transparent for application of optical measuring devices. A train of two-dimensional (2D) fixed dunes were installed on the bottom of the flume starting just downstream of the entrance on the flume. Dunes were prepared with well-sorted gravel particles. Experiments were carried out using four types of dunes with different lee slopes. The numerical model used was sediment simulation in intakes with multiblock option (SSIIM), which is an open-source computational fluid dynamics model. Both experimental observations and numerical results were in agreement about the influence of dune characteristics on the flow separation, with particular regard to dune height, particle size, angle o...

Evaluation of diffusion models for airborne nanoparticles transport and dispersion

The diffusion coefficient is a property that plays a significant role in the transport of airborne nanoparticles. However, there seems to be no general agreement in the literature on the most appropriate model to use for nanoparticle numerical simulations to be used in risk exposure assessments. This paper begins by presenting a brief review of some of the main models for small particles diffusion. A general dynamic equation for aerosol transport is briefly discussed next. Since the particle diffusion coefficient can be expressed in terms of a friction coefficient, three relationships are then presented and their influences on the friction and diffusion coefficients are considered for the particular case of TiO2 nanoparticles. Although, all the models studied here predict a decrease in the value of the diffusion coefficient with increasing particle diameter, some significant variations can be observed between the models. A specific diffusion model, chosen between those studied, is finally applied to estimate the purge time of airborne TiO2 nanoparticles in a simple closed space the size of a glovebox. It is shown that the sedimentation and the diffusion processes do not play a major role in the evaluation of the purge time.

Study of the near wake of a wind turbine in ABL flow using the actuator line method

The main goal of this study is the comparison of the near wake generated by an actuator line immersed in three different flows. A hybrid RANS-LES method is used to analyze the behaviour of the near wake in an atmospheric boundary layer. Then these results are compared to two idealized flow situations commonly found in the literature, such as irrotational flow and homogenous isotropic turbulence. In order to achieve this, a validation of the homogenous isotropic turbulence is conducted. The simulation results and the generated turbulence are compared with respect to the turbulence intensity and the integral length scale of the flow field. Finally to visualise the impact of the different flow conditions on the near wake of a wind turbine, the mean velocities, Reynolds stresses and energy spectra are compared.

Comparative CFD study of the effect of the presence of downstream turbines on upstream ones using a rotational speed control system

The effect of a downstream turbine on the production of a turbine located upstream of the latter is studied in this work. This is done through the use of two CFD simulation codes, namely OpenFOAM and EllipSys3D, which solve the Navier-Stokes equations in their incompressible form using a finite volume approach. In both EllipSys3D and Open Foam, the LES (Large Eddy Simulation) technique is used for modelling turbulence. The wind turbine rotors are modelled as actuator disks whose loading is determined through the use of tabulated airfoil data by applying the blade-element method. A generator torque controller is used in both simulation methods to ensure that the simulated turbines adapt, in terms of rotational velocity, to the inflow conditions they are submited to. Results from both simulation codes, although they differ slightly, show that the downstream turbine affects the upstream one when the spacing between the turbines is small. This is also suggested to be the case looking at measurements performed at the Lillgrund offshore wind farm, whose turbines are located unusually close to each other. However, for distances used in todays typical wind farms, this effect is shown by our calculations not to be significant.

Optimal distribution width for ALM in LES for (NEW) MEXICO experiment

The Actuator Line Method exists for more than a decade and has become a well established choice for simulating wind rotors in computational fluid dynamics. Numerous implementations exist and are used in the wind energy research community. A crucial parameter of this method is the distribution width used to distribute the blade forces in the domain in order to attain numerical stability but also to apply the force term over a length which makes physically sense. In this work the implementation is verified with results from the (NEW) MEXICO experiment and the optimal distribution width is found by setting the global torque as target variable.

Investigation of the viscous reconnection phenomenon of two vortex tubes through spectral simulations

This paper aims to shed further light on the viscous reconnection phenomenon. To this end, we propose a robust and efficient method in order to quantify the degree of reconnection of two vortex tubes. This method is used to compare the evolutions of two simple initial vortex configurations: orthogonal and antiparallel. For the antiparallel configuration, the proposed method is compared with alternative estimators and it is found to improve accuracy since it can account properly for the formation of looping structures inside the domain. This observation being new, the physical mechanism for the formation of those looping structures is discussed. For the orthogonal configuration, we report results from simulations that were performed at a much higher vortex Reynolds number (Re Γ ≡ circulation/viscosity = 104) and finer resolution (N 3 = 10243) than previously presented in the literature. The incompressible Navier-stokes equations are solved directly (Direct Numerical Simulation or DNS) using a Fourier pseudospectral algorithm with triply periodic boundary conditions. The associated zero-circulation constraint is circumvented by solving the governing equations in a proper rotating frame of reference. Using ideas similar to those behind our method to compute the degree of reconnection, we split the vorticity field into its reconnected and non-reconnected parts, which allows to create insightful visualizations of the evolving vortex topology. It also allows to detect regions in the vorticity field that are neither reconnected nor non-reconnected and thus must be associated to internal looping structures. Finally, the Reynolds number dependence of the reconnection time scale Trec is investigated in the range 500 ≤ Re Γ ≤ 10 000. For both initial configurations, the scaling is generally found to vary continuously as Re Γ is increased from Trec∼ReΓ−1 to Trec∼ReΓ−1/2, thus providing quantitative support for previous claims that the reconnection physics of two vortices should be similar regardless of their spatial arrangement.

Determination of friction factor over a pool reach by double-average method

Bed-form characteristics have been investigated because of the close relation to hydraulic friction. Pool is one of the most important bed forms in coarse-bed rivers. The form friction factor should be considered in determining many river engineering projects including stable channel design, the scour depth along bank protection and at bridges piers. This study was conducted in the coarse-bed, Saymareh river, in western Iran. Two reaches, with Froude number less than 1.0, were selected to investigate the contribution of the double-average method in determining friction factor, one with pool bed form, range of Reynolds numberfrom 1.325×10

Evaporation time of gasoline and diesel fuel droplets on a hot plate : the influence of fuel deposits

The objective of this paper is to present experimental results of multicomponent fuel droplets impinging on a hot surface in order to quantify the influence of fuel build-up deposits on the evaporation time. The experiments were conducted with gasoline and diesel fuels to first obtain curves of evaporation time as a function of plate temperature. Based on these curves the Nukiyama and Leindenfrost temperatures were identified. In a second step, the effect of fuel deposit on the droplet evaporation time was studied. Based on the above evaporation time curves, plate temperatures were chosen as to offer a similar evaporation time but at temperatures below and above the Nukiyama and Leindenfrost temperatures respectively. This was done in order to isolate the effect of fuel deposits from the different evaporation mechanisms. The evaporation of successive impinging droplets was then measured. The results hence obtained indeed showed that the fuel deposit has a different impact on the evaporation time according the evaporating mechanism or equivalently the plate temperature. For plate temperatures lower than the Nukiyama temperature, gasoline and diesel fuel droplets showed an increase of their evaporation time as the amount of successive impinging droplets increased. The trend was reversed for plate temperatures above the Leindenfrost temperature. A hypothesis for this latter case is that the fuel deposit disrupts the vapor layer supporting the droplet and therefore provides a greater heat flux to the evaporating droplet. Finally, droplet evaporation times as a function of plate temperature were measured with an initial fuel deposit covering the plate. These results in turn showed that the global thermal diffusivity and porosity of the surface are changed by the presence of the fuel deposit. The consequence of these property changes are then shown to have a direct and global impact on the fuel evaporation time curves.Copyright